AI is rewriting how developers build software but speed without guardrails creates risk. This session explores how AI coding assistants are accelerating development while introducing new cybersecurity vulnerabilities. Attendees will learn about prompt injection, insecure coding patterns, hallucinated libraries, and the “32% problem” of over-trusting AI outputs. The talk also covers practical strategies for embedding security into AI-assisted workflows and designing guardrails that protect innovation without limiting creativity.